The potential of artificial intelligence models in tourism education: Exam performance and ethical discussions

This study aimed to compare the exam performances of ChatGPT Plus and Google Gemini Advanced in tourism management, tourism marketing, and tourism economics courses with the exam performance of undergraduate students. One hundred fifty students studying at Harran University Faculty of Tourism and completing their education in these three courses were selected and included in the exams with artificial intelligence models. In the exam, 25 questions were created for each course by academicians who are experts in their fields. The results show that ChatGPT has the highest overall accuracy rate and the lowest number of wrong answers. ChatGPT gave 21 correct answers in the tourism economics exam, Google Gemini 18, and students 16.6. In the tourism marketing exam, ChatGPT 19 and Google Gemini 18, students gave 14.9 correct answers. In the tourism management exam, ChatGPT answered 22 questions correctly, Google Gemini answered 14 questions, and students answered 16.3 questions correctly. In addition, the questions were categorised as short, long, easy, medium difficulty, complex questions, negative sentences, and scenario questions. When the results were analysed, ChatGPT was more successful in all categories. Although artificial intelligence language models are more effective than undergraduate students in certain exam conditions, this study underscores the need for further research to optimise and validate the use of these technologies in education. In addition, as a result of the research, it is thought that artificial intelligence language models can play a transformative role in tourism education in the future. An important finding has also emerged that ensuring the ethical and practical use of artificial intelligence technologies in academic settings requires responsible integration, human oversight, and more validation studies.